Fei2411/real-api-pricing

What it solves

Many AI coding services charge a flat monthly fee rather than a per-token price, making it difficult to compare the true cost of different subscriptions. This project calculates the actual cost per million tokens for these plans, allowing users to see which services provide the best value for their money.

How it works

The project calculates a "real price" by dividing the monthly subscription fee by the total number of tokens the user is actually allowed to consume in a month. It then plots these prices against public AI model leaderboard scores on interactive charts. This helps visualize the "Pareto frontier," identifying plans that offer the highest performance for the lowest cost.

Who it’s for

This tool is designed for developers, researchers, and power users who want to make data-driven decisions when choosing between various AI coding subscriptions and metered API services.

Highlights

• Interactive Comparison: Visualizes the trade-off between subscription costs and model performance across multiple industry benchmarks.
• Standardized Metrics: Converts diverse pricing models and credit pools into a unified cost-per-million-token metric.
• CLI Support: Includes a TypeScript CLI to query the dataset, filter by company or plan, and integrate price checks directly into coding agent workflows.
• Transparent Data: Provides detailed evidence for every calculation, including confidence ratings and links to the specific decisions made for each plan.